Table 2.
Multiple model comparisons for presence and abundance of female Ae. aegypti, with daily and cumulative precipitation predictors.
| Model | Metric | Estimate |
|---|---|---|
| Logistic Presence intercept-only (null model) | AUC-ROC | 0.500 |
| Logistic Presence (days + 10 days cumulative precipitation) | AUC-ROC | 0.543 |
| Logistic Presence (days + 20 days cumulative precipitation) | AUC-ROC | 0.582 |
| Logistic Presence (Daily precipitation) | AUC-ROC | 0.656 |
| Poisson abundance intercept-only (null model) | RMSE | 4.850 |
| Poisson abundance (days + 10 days cumulative precipitation) | RMSE | 4.850 |
| Poisson abundance (days + 20 days cumulative precipitation) | RMSE | 4.832 |
| Poisson abundance (Daily precipitation) | RMSE | 4.844 |
Models for both presence and abundance are compared using each day’s daily precipitation as predictors, as well as cumulative precipitation with a 10 or 20 day horizon. Best-supported models are chosen through LASSO (least absolute shrinkage and selection operator) regressions and shown in bold. Logistic models are performed for presence and evaluated with area under the curve of the receiver-operating characteristic curve (AUC-ROC; with larger values corresponding to better model fits), while abundance models are fit to Poisson distributions and evaluated with root mean squared error (RMSE; with lower values corresponding to better model fits). Precipitation variables increase model performance for presence models, but abundance models all perform similarly, indicating poor model performance with only these variables.